对比6种多语言模型在尼泊尔语语音识别中的表现，发现语言亲缘性比参数量更重要。
Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition
- 统一微调协议下比较6种模型，涵盖自监督、自回归和混合架构。
- Whisper-Large与IndicWav2Vec在尼泊尔语上达到14.76%和14.89%的词错误率最优。
- 基于CTC的模型比Whisper快29倍，适合低延迟部署场景。
尽管多语言预训练模型宣称支持尼泊尔语，但缺乏在统一微调协议下的对照基准。本文在约165小时的OpenSLR SLR54尼泊尔语语料上，对六种预训练模型（XLSR-53、IndicWav2Vec、MMS-1B、Whisper-Medium、Whisper-Large-v3-Turbo、Conformer-Hi）进行相同预处理、数据划分、优化器和族系匹配的学习率调度下的微调。评估在三个独立测试集（OpenSLR、FLEURS、Common Voice）上的词错误率（WER）、字符错误率（CER）和实时因子（RTF）。Whisper-Large-v3-Turbo（14.76% WER）与IndicWav2Vec（14.89% WER）并列最优，尽管二者参数量相差9倍、预训练数据量相差40倍，表明语言家族亲缘性可替代规模以实现域内性能。基于CTC的解码器在同等准确率下比自回归Whisper快达29倍，颠覆了实际部署中对延迟的偏好。大规模多语言预训练（MMS-1B）在跨域评估中表现最稳健，FLEURS上仅增加12.55个百分点，说明规模提升鲁棒性而非峰值性能。该基准为尼泊尔语语音识别提供了首个标准化、多模型、效率感知的参考指标。
原文摘要 · Abstract (English)
Multilingual pretrained models nominally support Nepali, yet no controlled benchmark has compared them under a single fine-tuning protocol. We fine-tune six pretrained models (XLSR-53, IndicWav2Vec, MMS-1B, Whisper-Medium, Whisper-Large-v3-Turbo, and Conformer-Hi) spanning CTC self-supervised, autoregressive encoder-decoder, and hybrid Conformer-CTC architectures, on the OpenSLR SLR54 Nepali corpus (~165 hours) using identical preprocessing, splits, optimizer, and family-matched learning-rate schedules. We evaluate Word Error Rate (WER), Character Error Rate (CER), and Real-Time Factor (RTF) on three independent test sets (OpenSLR, FLEURS, Common Voice). Whisper-Large-v3-Turbo (14.76% WER) and IndicWav2Vec (14.89% WER) tie at the top despite a 9x parameter gap and 40x pretraining-data gap, providing direct empirical evidence that language-family proximity in pretraining can substitute for raw scale for in-domain Nepali. CTC decoders run up to 29x faster than autoregressive Whisper at the same accuracy, flipping the practical deployment preference toward CTC under any latency budget. Massively multilingual pretraining (MMS-1B) yields the smallest out-of-domain degradation on FLEURS (+12.55 pp), indicating that scale buys robustness rather than peak in-domain accuracy. The resulting benchmark provides the first standardized, multi-model, efficiency-aware reference numbers for Nepali ASR.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。